Wolfram Mathematica

Q: Default behaviour of Hash[expr] and hashing in different versions of Mathematica

9:34 AM

@SjoerdC.deVries When you get any information from support@WRI, could you add it as answer here?

In different versions of Mathematica, Wolfram silently changed the behaviour of Hash when the algorithm is not specified explicitly Hash[1] (* 6568131406215528669 (Version 10.1) *) Hash[1] (* 4371187653775642860 (Version 9.0.1) *) Hash[1] (* 1742717557 (Version 8.0.4) *) This is a serious vi...

functions bugs

blochwave

@kirma make it community wiki?

3 hours later…

Sjoerd C. de Vries

12:17 PM

@halirutan will do

12:40 PM

Has anybody else experienced strange multi-kernel busylooping on 10.1? I seem to experience that, starting from something that has nothing to do with my Mma usage (like opening laptop lid). I think I have seen similar behaviour on some earlier versions, but for a while...

xslittlegrass

1:37 PM

@kirma I have some issue with the remote kernel on 10.1. Mathematica frontend (and sometimes a kernel) uses nearly 100% CPU even no task running. I'm not sure whether it's related or not.

@kirma chat.stackexchange.com/transcript/message/20889849#20889849

@xslittlegrass Hmmh. This is entirely 10.1, running only on the laptop. Computing kernels occassionally start busylooping, and if I have several started, all of them.

xslittlegrass

2:03 PM

@kirma OK then my maybe different. But this kind of bug sometimes is annoying because it is difficult to pinpoint when it will happen.

@xslittlegrass It might be the same too, though.

A: Do we need a tag for documentation bugs?

2:42 PM

Interesting. I wonder if my math makes sense, it would seem that mean lengths of lattice-free unit-step random walks are actually rational numbers.

At least in 2D Euclidean geometry.

1 hour later…

Szabolcs

4:01 PM

I think the bugs tag really shouldn't be used for documentation problems. There's so much potential for misunderstanding. The tag is already sensitive enough, stirs up feelings easily, gives ways to rants, and generally requires more moderation than usual. There should be clear and strict rules for when it can be used.

Given how much more trivial a documentation inaccuracy is than a real software bug, I believe this tag should be reserved for real software bugs only.

Meta post:

I think we should not use the bugs tag for inaccuracies in the documentation. It should be reserved for situation when there clearly is a bug in the software. Reasons: The bugs tag is already a sensitive and high maintenance tag. It should serve a useful purpose, and we must not allow it to b...

notebooks string-manipulation

@Szabolcs What about this? Is this really a documentation bug and do we really expect the documentation to tell us something like: "Hash is a surprising function which will change its behaviour in every version and you won't be able to find out what exactly we are doing in default mode.."?

2 hours later…

Sjoerd C. de Vries

5:59 PM

@halirutan I just saw that the behavior of Hash has been discussed as early as 2010...

Liam William

Q: Split at linebreak

How do I split the following file at linebreaks. RowBox[{"$CommandLines", "[", RowBox[{"[", "3", "]"}], "]"}], ",", "\"\<Text\>\""}], "]"}], ",", "\"\<\\n\>\""}], "]"}], ",", "3"}], "]"}], ",", "\"\< \>\""}], "]"}]}]}], "]"}], ";", "\[IndentingNewLine]", ...

5 hours later…

10:46 PM

@MichaelE2 Are you still around?

11:03 PM

@halirutan Yep

@MichaelE2 I have the feeling of a deja vu because I think I thought about this context problem, but I don't clearly remember. Can you make any statement, why symbol resolution doesn't work as it worked in earlier versions? Simple example:

x = 1
test`x = 3

Begin["test`"]
x
End[]

@MichaelE2 Why does it use the global definition of x, when we are in the test context?

The documentation to $ContextPath clearly says, that the current context is searched first:

> $ContextPath is a global variable that gives a list of contexts, after $Context, to search in trying to find a symbol that has been entered.

@MichaelE2 And then we have this (please Quit[])

x = 1
Begin["test`"]
x = 3;
Context[x]
x
End[]

Here, the symbol definition x=3 is not done in the current $Context but the existing definition for Global`x is changed.

@halirutan I do not know. It would appear that $ContextPath is searched first.

@MichaelE2 Hmm.. this is all pretty weird. Thanks nevertheless.

11:24 PM

@halirutan It's different (of course?) if you replace Begin/End with BeginPackage etc. And you're welcome.

@MichaelE2 Yes, it is different because BeginPackage resets the $ContextPath.

@halirutan That was what I meant by "(of course?)" -- the question mark indicates that it shouldn't be different if $Context is searched first.

@MichaelE2 Ah, OK.

@halirutan Quit[] and then run this, and then run it again without quitting:

Begin["test`"]
x = 3;
Begin["foo`"]
x = 5
Context[x]
x
End[]
x
End[]
x

The last x creates Global`x, which messes up the second evaluation. Apparently, Global` is a special case.

@MichaelE2 The problematic part is that you mention x in the very end which creates the symbol in the global context :-)

After that, everything is weird

@MichaelE2 The thing is that the documentation doesn't mention anything of this throughout everything I have seen so far.

@MichaelE2 Oh, I found something that mentions it is the other way around:

> Since $Context is searched after $ContextPath, you can think of it as having "." appended to the file search path.

from reference.wolfram.com/language/tutorial/Contexts.html

11:42 PM

@halirutan Yes, I put the x at the end to test that. I found that tutorial.

@halirutan I guess the $ContextPath doc page is the typo.

@MichaelE2 I don't know. The sentence sounds (with its comma insertion) a bit strange to me. A native speaker could tell whether the sentence could be understood the other way arround

Writing up an answer to the question in question..

@halirutan As a native speaker, a phrase set off by commas is parenthetical. This one reminds the reader that the contexts in $ContextPath are searched in order after $Context is searched. I don't see how to interpret it any other way. "Followed by $Context" would be the way to convey the reverse sense.

@MichaelE2 OK, then we understood this in the same way.